+(,121/,1(
Citation: 105 Colum. L. Rev. 1124 2005
Content downloaded/printed from
HeinOnline (http://heinonline.org)
Wed Nov 17 15:52:44 2010
-- Your use of this HeinOnline PDF indicates your acceptance
of HeinOnline's Terms and Conditions of the license
agreement available at http://heinonline.org/HOL/License
-- The search text of this PDF is generated from
uncorrected OCR text.
-- To obtain permission to use this article beyond the scope
of your HeinOnline license, please use:
https://www.copyright.com/ccc/basicSearch.do?
&operation=go&searchType=0
&lastSearch=simple&all=on&titleOrStdNo=0010-1958
SENTENCING DECISIONS: MATCHING THE
DECISIONMAKER TO THE DECISION NATURE
Paul H. Robinson* & BarbaraA. Spellman**
The present sentencing debate focuses on which decisionmaker is best
suited to make the sentencing decision. Competing positions in this debate
typically view the sentencing decision as monolithic, preferring one decisionmaker over all the others. A monolithic view of the decision unnecessarily invites poor decisionmaking. The sentencing decision is properly viewed
as a series of distinct decisions, each of which can best be performed by a
decisionmakerwith certain qualities. This Essay demonstrates how a system
of optimal decisionmakingmight be constructed-by sortingout the different
attributes called for by the distinct aspects of the sentencing decision and
matching them to the strengths and weaknesses of each potential
decisionmaker.
1125
1128
INTRODUCTION ....................................................
I.
THE DECISION: MONOLITHIC OR MULTIFARIOUS? ...........
A. Decision 1: Policymaking-Setting the Purposes,
Goals, and Priorities .................................
B. Decision 2: Rule Articulation-Translating General
Policy into Articulable Rules .........................
C. Decision 3: Factfinding-Re-creating Past Events .....
D. Decision 4: Judgment-Making-Expressing Normative
Judgm ents ...........................................
E. Decision 5: Determining Punishment AmountApplying the Case Findings Determined in Decisions 3
and 4 to the Articulated Rules Determined in
D ecision 2 ...........................................
F. Decision 6: Determining Punishment MethodTranslating the Decision 5 Conclusions Concerning
Punishment Amount into a Specific Sentence ........
II.
III.
THE DECISIONMAKERS .......................................
MATCHING THE DECISIONMAKER TO THE DECISION NATURE
A. Decision 1: Policyrmaking ............................
B. Decision 2: Rule Articulation ........................
C. Decision 3: Factfinding ..............................
1129
1129
1130
1130
1130
1131
1132
. 1132
1132
1136
1138
1.
Jury (VersusJudge) Decisionmaking Literature ...
1138
2.
Other Relevant Social Science Literature .........
a. Training and Experience .....................
1140
1142
* Colin S. Diver Distinguished Professor of Law, University of Pennsylvania. The
authors thank John Darley and Reid Hastie for their insights and Elizabeth Gerlach and
Michael Areinoff for their research assistance.
** J.D., Ph.D., Associate Professor of Psychology, University of Virginia.
1124
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
Individual Differences Between Judges and
Juries ........................................
c. Groups Versus Individuals ....................
d. Selection and Perception .....................
D. Decision 4: Judgment-Making ........................
E. Decision 5: Determining Punishment Amount .......
F. Decision 6: Determining Punishment Method ........
1125
b.
IV.
IMPLICATIONS FOR SENTENCING REFORMS ...................
A. Mandatory Minimum Sentences ......................
B. Federal Sentencing Reform Act of 1984 ..............
C. The Feeney Amendment to the Federal PROTECT
Act: As Proposed, as Enacted .........................
D. Blakely v. Washington .................................
E. Jury Sentencing ......................................
F. State Sentencing Guidelines Systems ..................
CONCLUSION
....................................................
1142
1143
1145
1146
1148
1149
1151
1152
1152
1154
1156
1158
1159
1161
INTRODUCTION
A large part of the present sentencing debate concerns the question
of who should make the sentencing decision. Traditional practice gives
judges broad discretion to impose any sentence less than the statutory
maximum, which typically is sufficiently high as to present little practical
restriction. In many systems, however, judicial sentencing discretion is a
bit of an illusion because the parole board, making its own independent
assessment of the case upon the offender's arrival at prison, sets the presumptive release date.' Mandatory minimums shift some of what was traditionally a judicial sentencing decision to the legislature. 2 Guidelines
shift much of this decisionmaking to a sentencing commission, especially
under binding guidelines such as those of the U.S. Sentencing Commission (Commission),3 although the recent PROTECT Act shifts some of it
back to Congress 4 and would have shifted quite a bit more but for intense
1. See, e.g., Mass. Gen. Laws Ann. ch. 27, § 5 (West 2002) (authorizing the
Massachusetts Parole Board to determine which prisoners may be released, when, and
under what conditions); Pa. Stat. Ann. tit. 61, § 331.17 (West 1999) (granting the
Pennsylvania Board of Probation and Parole exclusive authority to parole and reparole
offenders who have been sentenced for a maximum period of at least two years).
2. See, e.g., StephenJ. Schulhofer, Rethinking Mandatory Minimums, 28 Wake Forest
L. Rev. 199, 201-02 (1993) (discussing federal mandatory minimum sentence provisions).
3. See Sentencing Reform Act of 1984, Pub. L. No. 98-473, 98 Stat. 1987 (codified as
amended in scattered sections of 18 U.S.C. and 28 U.S.C. (2000)); see U.S. Sentencing
Guidelines Manual § 1AI.1, cmt. background (2003).
4. See Prosecutorial Remedies and Other Tools to end the Exploitation of Children
Today Act of 2003 (PROTECT Act), Pub. L. No. 108-21, 117 Stat. 650 (codified as
amended in scattered sections of 18 U.S.C. and 28 U.S.C.) (making various changes to the
U.S. Sentencing Guidelines without consulting the Commission, including restricting
downward departures in certain cases by directly rewriting guidelines, ordering the
Commission to further reduce downward departures, restricting the Commission's ability
1126
COLUMBIA LAW REVIEW
[Vol. 105:1124
lobbying against the legislation. 5 Some writers see the Supreme Court's
Apprendi6 and Blakely7 decisions as demanding a shift of some of the sentence decisionmaking to the jury. 8 Some jurisdictions presently have
forms of explicit jury sentencing. 9
Competing positions in the sentencing debate often reduce to a
writer's preference for one decisionmaker over another. Perhaps not surprisingly, legislators commonly prefer legislatures.' 0 Judges also have
strong supporters. One sentencing scholar recently argued that judges
to amend the guidelines to include new grounds for downward departures, increasing
reporting requirements, and altering appellate review).
5. See, e.g., Alan Vinegrad, The New Federal Sentencing Law, 15 Fed. Sent'g Rep.
310, 315-16 (2003) (discussing last-minute changes to the Feeney Amendment to the
PROTECT Act in response to criticisms).
6. Apprendi v. New Jersey, 530 U.S. 466, 490 (2000) (requiring that any fact, other
than a prior conviction, that increases a sentence beyond the statutory maximum be
submitted to the jury and be proven beyond a reasonable doubt).
7. Blakely v. Washington, 124 S. Ct. 2531, 2537-38 (2004) (holding that defendant's
sentence under a state guidelines system violated his Sixth Amendment right to ajury trial
because the sentencing judge increased his sentence above the guideline range based on
an aggravating factor not found by a jury nor admitted by defendant).
8. See, e.g., Morris B. Hoffman, The Case for Jury Sentencing, 52 Duke L.J. 951,
955-56 (2003) (suggesting that Supreme Court decisions, including Apprendi, have opened
up the debate about whether noncapital sentencing should remain in the hands of judges
rather than juries); NancyJ. King & Susan R. Klein, Beyond Blakely, 16 Fed. Sent'g Rep.
316, 322-23 (2004) (discussing how Blakely shifts factfinding part of the sentencing
decision to the jury).
9. Six states-Virginia, Kentucky, Arkansas, Missouri, Texas, and Oklahomacurrently have jury sentencing in felony, noncapital cases. See NancyJ. King & Rosevelt L.
Noble, Felony Jury Sentencing in Practice: A Three-State Study, 57 Vand. L. Rev. 885, 886
(2004). In each of these jurisdictions, judges instruct juries on the statutory range of
punishment and juries select a sentence within that range. Id. Virginia, Kentucky, and
Arkansas each bifurcate jury trials to provide for separate guilt determination and
sentencing phases. Id. at 892. In those states, the judge may reduce the jury's sentence,
but may only raise the sentence if necessary to comply with mandatory minimum statutes.
Id. Virginia and Arkansas both have voluntary judicial sentencing guidelines; however,
juries are not instructed on these guidelines and instead select a sentence from within the
statutory range, as in other states. Id. at 893. Kentucky and Arkansas allow juries to hear
evidence as to eligibility for parole, but not evidence as to how often parole release actually
occurs. Id. at 893-94.
10. Witness, for example, the legislative resistance to withdrawing mandatory
minimum sentences even after a sentencing guideline system is in place to control judicial
discretion. The United States Sentencing Commission issued a report calling on Congress
to repeal mandatory minimums because they were no longer necessary and were
interfering with the rational operation of federal sentencing guidelines. U.S. Sentencing
Comm'n, Special Report to the Congress: Mandatory Minimum Penalties in the Federal
Criminal Justice System (1991), available at http://www.ussc.gov/r_congress/MANMIN.
pdf (on file with the Columbia Law Review). "Commentators and judges from across the
political spectrum have called for elimination of mandatory penalties." Marc L. Miller,
Domination & Dissatisfaction: Prosecutors as Sentencers, 56 Stan. L. Rev. 1211, 1265
(2004) (citing examples). Yet Congress has refused to withdraw mandatory minimums
and, indeed, has sought to take sentencing authority back from the federal sentencing
commission. See infra Part IV.C.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1127
are particularly "institutionally well suited" to the task, better suited than
the legislature or sentencing commissions because of a "combination of
perspective, process and politics." 1 There is also much support for sentencing commissions.
The [Sentencing Reform Act of 1984] was enacted against a history... of racial, geographical and other unfair disparities in
sentencing.... IT]he task of harmonizing sentencing policies
was deliberately placed in the hands of an independent Sentencing Commission. The guidelines as originally conceived were
about fairness, consistency, predictability, reasoned discretion
and minimizing the role of congressional politics and the ideol12
ogy of the individual judge in sentencing.
Parole boards also have their strong supporters. 13 And there is growing support for jury decisionmaking.
11. Douglas A. Berman, A Common Law for This Age of Federal Sentencing: The
Opportunity and Need for Judicial Lawmaking, 11 Stan. L. & Pol'y Rev. 93, 110 (1999).
Professor Berman goes on to say:
From their case-specific perspective, judges individually consider and pass
sentencing judgment on the real persons who actually commit offenses....
The process through which federal judges make sentencing decisions also
seems especially well suited to deliberative, balanced and less punitive sentencing
lawmaking....
... At the same time, while their position allows them to avoid the (usually
punitive) pressures of politics, federal judges must directly confront the (often
sobering) considerations of costs. Each sentencing provides judges a firsthand
view of both the financial costs and the human costs that severe sentencing laws
produce.
Id. (footnotes omitted).
12. Blakely v. Washington and the Future of the Federal Sentencing Guidelines:
Hearing Before the Senate Comm. on the Judiciary, 108th Cong. (July 13, 2004), available
at 2004 WL 1638424 (statement of Sen. PatrickJ. Leahy, Ranking Member, Senate Comm.
on the Judiciary). In the same hearings, Bill Mercer, United States Attorney, District of
Montana, similarly testified:
Congress unified under the common recognition that unstructured criminal
sentencing had evolved into a vehicle for disparity in actual punishment that
simply could not be justified and uncertainty in sentencing that was contributing
to intolerable levels of crime. Offenders with similar criminal histories who
committed similar offenses often received and served substantially different
sentences.... Under the Sentencing Reform Act of 1984, offenders with similar
criminal histories who commit similar offenses receive similar sentences, because
sentencing courts are directed to evaluate specific, enumerated factors in the
guidelines and to engage in rigorous and appealable fact finding to determine
whether these factors are present in each case. The sentences handed down
under the guidelines have been predictable.
Id. See generally Michael Tonry, The Success of Judge Frankel's Sentencing Commission,
64 U. Colo. L. Rev. 713 (1993) [hereinafter Tonry, Judge Frankel]; A Symposium on
Sentencing Reform in the States, 64 U. Colo. L. Rev. 645 (1993).
13. Parole boards, among other things, provide an opportunity to bring consistency to
the often disparate sentencing judgments made by individual sentencing judges. This
attribute has often been the justification for the continued existence of parole boardseven after they started making their parole decisions at the beginning of prison terms
(thus using the same information available to the sentencing judge) instead of the end.
1128
COLUMBIA LAW REVIEW
[Vol. 105:1124
Because of their deliberative capacity and democratic makeup,
juries are better situated than other political institutions to perform the sensitive tasks of deciding between contested sentencing goals and applying the law with due regard for the individual circumstances of each offender. In these respects, juries
have a significant advantage over legislatures, agencies, and
judges.14
While these statements raise valid points, a more accurate view is that
each decisionmaker has both advantages and disadvantages as a sentencer. Most importantly, each of these advantages and disadvantages
has a different impact on different aspects of the sentencing decision.
This Essay argues that the sentencing decision is properly viewed as a
series of decisions-each of which is importantly different from the
others and each of which can best be performed by a decisionmaker with
certain qualities. To prefer one decisionmaker over another, without distinguishing among the different aspects of the sentencing decision-as
the commentators quoted above seem to do-is to invite, unnecessarily,
poor decisionmaking. One might construct a system of optimal sentencing decisionmaking by this exercise: (i) sort out the aspects of the sentencing decision that call for different kinds of decisionmaking attributes,
(ii) sort out the strengths and weaknesses of each potential decisionmaker, and (iii) match the latter to the former.
The structure of this Essay follows this analytic path. The analysis is
meant to be primarily illustrative. No doubt social science research more
focused on these decisionmaking questions could produce more credible
conclusions than those offered here. It is the inquiry and the illustrated
process, rather than its specific conclusions, that are offered as most
useful.
I. THE DECISION: MONOLITHIC OR MULTIFARIOUS?
Imagine a world in which a group could, with complete discretion,
decide the sentences of each offender. This exercise is not too different
from the decision facing the sentencing judge with unlimited sentencing
discretion that was standard not too long ago. Imagine further that this
decisionmaking group is populated only with capable, thoughtful, educated, and experienced people. What would the group's ideal decision
process look like? One could imagine this as a rather monolithic decision: known facts in, punishment decision out. But it would not take
long-perhaps after only a few cases, some similar and some differentfor this group of thoughtful sentencers to conclude that the sentencing
decision is multifaceted rather than monolithic. The need for a sentence
to serve a rational purpose, the need for some uniformity in application
of punishment, the need for some transparency in the sentencing pro14. Jenia Iontcheva, Jury Sentencing as Democratic Practice, 89 Va. L. Rev. 311, 350
(2003).
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1129
cess, the need to create public confidence in unbiased application, and a
variety of other interests are likely to quickly suggest to our sentencing
group that the sentencing decision is more than facts in and punishment
out, but rather something that necessarily has a number of distinct
pieces. There are at least six distinct decisions to be made.
A. Decision 1: Policymaking-Settingthe Purposes, Goals, and Priorities
What are we trying to achieve in criminal sentencing? What purpose
are we trying to advance: general deterrence, special deterrence, rehabilitation, incapacitation, just deserts, reinforcing criminal law's moral credibility, reconfirming the criminal law's prohibitions, other goals, or some
combination of these? To make these choices, one must understand the
strengths and weaknesses, the advantages and disadvantages, of each alternative sentencing goal. One must also consider what other interests
might properly constrain the system's move toward these goals. Are there
notions of fairness that must be reflected in its procedures? Are there
notions of justice and equality of treatment that insist on uniformity in
the treatment of cases that are similar in meaningful ways? The answers
to these questions lie in a variety of disciplines, including criminology,
psychology, moral philosophy, criminal law theory, and political science.
The answers also require the most fundamental of value judgments,
which essentially define the very nature of a society and its priorities.
B. Decision 2: Rule Articulation-TranslatingGeneral Policy into Articulable
Rules
To produce a sentence, the broad policy decisions described above
must be reduced to specific rules. 15 This is especially true if there is to be
a possibility of some degree of uniformity among similar cases, but would
be needed even if uniformity were not a concern. For example, if it is
determined that sentencing ought to maximize general deterrent effect,
that policy must be operationalized into specific rules that will achieve
that result. That articulation of rules requires considerable knowledge
and some analysis of how sentences should be set to efficiently optimize
general deterrence. Must there be greater punishment for more serious
harm caused? Greater punishment for offenses of intention than for offenses of recklessness? Greater punishment for offenses with lower detection rates? If so, under what circumstances? Further, if these factors are
to be relied upon, the articulating rules cannot be fully operationalized
15. In a system of completely discretionary sentencing by a single judge, the judge
might just do this in her head. In any other kind of system, the general policy goals must
be reduced to some kind of articulated sentencing rules, guidelines, or shared practices.
Few would argue that there is virtue in not articulating sentencing rules, guidelines,
preferences, or whatever else one might term these articulations. Rather, the present
disputes focus on first, which potential decisionmaker should undertake the task; and
second, whether there are countervailing interests (like maximizing discretion) that
suggest this ideal is one that must be compromised.
1130
COLUMBIA LAW REVIEW
[Vol. 105:1124
without also providing rules that define what is a "more serious harm," or
an "intention," and so forth.
C. Decision 3: Factfinding-Re-creatingPast Events
Applying the decision rules to the case at hand requires, first, an
understanding of what actually happened in the case. The decisionmaker must essentially re-create, as best she can, the past events leading up to the commission of the offense (and occasionally after), with
special focus on those facts that might be relevant under the decision
rules developed in Decision 2. Who did what, when, and how? Often
"why?" can also be important, as well as the particularly difficult "with
what culpable state of mind?"
D. Decision 4: Judgment-Making-ExpressingNormativeJudgments
The re-creation of past events is only part of the inquiry needed to
understand the case at hand. An assessment of punishment commonly
requires decisions that are less a matter of finding facts than a matter of
making normative judgments. Was the extent of the offender's distortion
of perception or impairment of control at the time of his offense sufficient to affect his sentence? If so, how much? Was the strength of the
causal connection between the offender's conduct and the harmful result
enough such that the resulting harm should affect the sentence? If so,
how much? Was the extent of the offender's assistance to the offense
conduct of another perpetrator enough to affect the sentence? If so, how
much? Did the offender's conduct go far enough to constitute an attempt? If so, to what extent should this determination affect his
sentence?
These sorts of questions cannot be resolved by simple fact~finding.
They call for normative judgments to be made upon the factual circumstances re-created in answer to Decision 3. Even a video recording of
exactly what happened, with a truthful narration track by each participant as to what he was thinking at the time, would leave the decisionmaker only at the starting point in trying to give an answer to these
16
issues ofjudgment.
E. Decision 5: Determining Punishment Amount-Applying the Case Findings
Determined in Decisions 3 and 4 to the Articulated Rules Determined
in Decision 2
Once the facts are found (Decision 3) and the factors judged (Decision 4), the governing decision rules must be applied (Decision 2) to
determine the amount of punishment to be imposed-such amount be16. Of course, whether such normative judgments are relevant may depend upon
those principles that one has set as guiding the distribution of punishment in Decision 1.
Certainly, if notions ofjustice are to have any relevance, normative judgments cannot be
avoided.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1131
ing a function of the purpose sought to be advanced (Decision 1). If the
distributive principle is desert, for example, the measure of amount will
be a product of the person's overall blameworthiness judged in light of
all relevant factors-the seriousness of the offense, the actor's level of
culpable state of mind at the time of the offense, the capacity of the actor
to avoid the offense conduct, and so forth. (The continuum of potential
blameworthiness will lie upon the continuum of punishment amount that
the society will permit.) If the distributive principle is instead the incapacitation of dangerous persons, the decision regarding the amount of
punishment will look to the length of time for which the offender is predicted to be sufficiently dangerous to justify government intervention. If
the distributive principle is deterrence, the amount decision will be a
function of the degree of punitive bite that is needed to deter a future
violation by this offender (special deterrence) or other potential offenders (general deterrence).
F. Decision 6: DeterminingPunishment Method-Translatingthe Decision 5
Conclusions ConcerningPunishment Amount into a Specific Sentence
The issue of how much punishment is to be imposed (Decision 5) is
distinct from the issue of how that punishment is to be imposed. 17 A
given amount of punishment could be imposed through any number of
different methods or combination of methods-imprisonment, house arrest, fine, weekend jail, community service, drug treatment, and so forth.
In a desert-based system or a deterrence-based system-which attempt to
modulate the amount of suffering either inflicted or threatened, respectively-the greater the "bite" of the punishment method the more punishment "credit" it should contribute to satisfying the total punishment
amount called for. In a system aimed at incapacitating the dangerous,
the decision among the various methods of restraint-a term of imprisonment, home detention, regular drug testing and monitoring, etc.-will
depend upon which method most effectively serves the protective function during Decision 5's predicted period of dangerousness, ideally minimizing the extent of liberty intrusion and cost consistent with societal
protection.
Whatever the sentencing purpose, any variety of sanctioning methods is possible. The choice of punishment method can be highly complex, requiring a good deal of information both about alternative punish17. Different methods of punishment (Decision 6) might be used to satisfy the same
amount of punishment (Decision 5). For example, a given period of imprisonment may
be the punitive equivalent of a longer period of home detention. This type of calculation
depends upon determining the proper ratios of punitive "bite" among different
punishment methods. See Robert E. Harlow, John M. Darley & Paul H. Robinson, The
Severity of Intermediate Penal Sanctions: A Psychophysical Scaling Approach for
Obtaining Community Perceptions, 11 J. Quantitative Criminology 71, 86-90 (1995)
(discussing results of empirical study measuring perceptions of relative severity of different
punishment methods).
1132
COLUMBIA LAW REVIEW
[Vol. 105:1124
ment methods and their effects and about the offenders and their
present circumstances.
II. THE DECISIONMAMERS
The five decisionmakers most commonly having a role in existing
American criminal sentencing systems-judges, juries, legislatures, sentencing commissions, and parole boards's-are different from one another in some rather obvious ways. Judges are different from the others
in that they act individually. Juries are different from most other decisionmakers because they lack the expertise that comes with special training or full-time participation. Legislatures are fairly large groups, in
which most members will not have expertise, but some will. Sentencing
commissions and parole boards are smaller in size, and most members
are likely to have expertise. The two groups, however, are traditionally
different from one another in that the sentencing commissions only produce guiding rules while the parole boards also apply rules to individual
cases. Also, parole boards typically involve themselves only in sentences
of imprisonment, while sentencing commissions commonly influence all
forms of sanction. 19
III. MATCHING THE DECISIONMAKER TO THE DECISION NATURE
For each of the six kinds of sentencing decisions that must be made,
consider the suitability of each of the five potential decisionmakers in
making such a decision.
A. Decision 1: Policymaking
The easiest of the five potential decisionmakers to exclude from policymaking would seem to be the jury. Jurors rarely have the knowledge
necessary to make informed policy decisions, and it is unrealistic to expect that they could be given that knowledge, short of sending them all to
college and graduate school. (See cell ID in the table below, which summarizes the conclusions of this Part.)
Judges also seem a poor choice. No doubt they have a good deal of
education, but they typically lack the breadth of education, such as a social science research background, and the research resources required to
18. For practical reasons we focus here on the five most obvious decisionmakers as
they commonly function today. One can imagine constructing an entirely different
decisionmaking body, such as a body of professionals like the membership of a sentencing
commission that not only sets policy but also adjudicates individual cases as a parole board
would. An example of such a decisionmaker is the Pennsylvania Sexual Offender
Assessment Board. See 18 Pa. Cons. Stat. Ann. § 9795.4 (West 2002) (authorizing the State
Sexual Offender Assessment Board to conduct assessments of individual offenders and
establish standards governing individual assessments).
19. See supra notes 1, 13. Early parole boards made release decisions only at the back
end of imprisonment, but modern boards typically set presumptive release dates at its
start-essentially acting as resentencers after the more public judicial courtroom.
20051
MATCHING DECISIONMAKER TO DECISION NATURE
1133
make policy decisions in a fully informed manner. Further, to the extent
that much of this decisionmaking requires a balancing of fundamental
societal values-how important is crime reduction versus uniformity in
application versus doing justice, and so forth-no single individual is
likely to be able to capture and represent the range of views within the
society. Perhaps most damaging for judges as policymakers is the need
for such decisions, whatever they are, to be applied to all offenders. An
offender's sentence ought to depend on the facts of his case, not on the
particular philosophy of the sentencing judge he happens to draw. Studies have repeatedly confirmed the wide variety of sentencing philosophies
among judges, as well as the effect such differences in philosophy have in
producing different sentences for identically situated offenders. 2 0 For all
these reasons, it is perhaps astonishing to think that the complete and
unguided judicial discretion commonly permitted in criminal sentencing
until the past few decades left policymaking decisions to judges.2 1 (See
cell lE.)
Sentencing commissions and parole boards both avoid the difficulties of having a sentencing judge as decisionmaker. They either can have
among their members the expertise required or can have the institu20. See, e.g., John Hogarth, Sentencing as a Human Process (1971). In a
monumental study of the attitudes and sentencing decisions ofjudges and others, Hogarth
found that "[w]ith the use of multiple regression methods, it was possible to demonstrate
that the sentencing behaviour of magistrates (for every offence studied) was significantly
associated with the attitudes of the magistrates concerned." Id. at 163. For example,
"[tihe frequent use of suspended sentence without probation in indictable cases is
associated with a lack of concern for abstract justice, tolerance of deviance, the view that
punishment does not correct the offender, and modernism in outlook. Taken together
these attitudes represent a liberal, non-punitive sentencing philosophy." Id. at 161.
Additionally,
[M] agistrates who frequently use institutional sentences [prison] for theft tend to
be traditional in outlook and rather intolerant in their attitudes to social
deviance. Once again, this is as one would expect. Theft is not often a
particularly serious offence. A tolerant magistrate or a magistrate who identifies
with modern doctrines would not be likely to use imprisonment for cases of this
kind.
Id. at 162-63.
21. It could be argued that wide discretion in the individual case is necessary in this
context because different underlying goals of sentencing should apply in different
situations: for example, general deterrence in one case and rehabilitation in another. But
this view fails to appreciate that such a system-allowing the policy decisionmaker to shift
among sentencing purposes for different cases-creates a dangerous facade. While the
resulting decisions may seem principled-they can be justified as serving the cited
purpose-without a principle that defines the interrelation among the competing
purposes, a decisionmaker may shift among purposes without reason or predictability.
That is, different purposes commonly will call for different results in a given case, and
without a defined interrelation among the different purposes, a decisionmaker can
determine a desired result and then, consciously or unconsciously, work backwards to cite
the sentencing purpose that reaches the desired result. For more details on this problem,
see Paul H. Robinson, Criminal Law 23-25 (1997); Paul H. Robinson, Hybrid Principles
for the Distribution of Criminal Sanctions, 82 Nw. U. L. Rev. 19, 19-22 (1987).
1134
COLUMBIA LAW REVIEW
[Vol. 105:1124
tional resources to consult or employ a wide range of experts in a variety
of disciplines. As a group, they have a better chance of representing the
range of different points of view on matters of value judgment. And, because of their institutional status, their policymaking decisions have effect
in more than just an individual judge's cases, thereby avoiding the disparity problem inherent in judicial policymaking. One difference between a
sentencing commission and a parole board may be that the latter, in addition to its policymaking role, typically also has the additional and quite
different role of applying those policies in individual cases. One could
speculate that this additional individual-case role might have advantages
or disadvantages for a policymaker. However, as noted, the sentencing
commission has a key advantage that would seem to moot the issue: Only
the commission is in a position to set policy for all sentences, not just a
subgroup of cases that are previously determined to merit a term of imprisonment. (See cells 1B & 1C.)
The legislature as policymaker would ensure reliance upon the same
policies in all sentencings as effectively as policymaking by a sentencing
commission. On the other hand, the legislature as policymaker seems
slightly less attractive than a sentencing commission in some respects. To
the extent that changing conditions require a reevaluation of policies,
continual supervision and adjustment are more realistically performed by
a commission. Further, a commission may have an advantage on matters
requiring expertise. The legislators themselves are not likely to have the
expertise needed, and, while they could hire experts, it might be overly
optimistic to think that they would or could spend the time and have the
expertise necessary to educate themselves to the extent required for responsible policymaking.
But because the legislature is more representative of the society than
a sentencing commission could be, the legislature might be a better
choice to make what are essentially broad value judgments in choosing
among competing purposes and in weighing competing interests, such as
choosing among the often conflicting interests of justice, crime control,
uniformity in application, and procedural fairness. One might conclude,
then, that the best allocation of the policymaking decision would be for
the legislature to have primary authority, but where there is a need for
expertise, flexibility, or continuing reevaluation and revision, the legislature would delegate that authority to a sentencing commission.
Another argument for a split of authority between legislatures and
sentencing commissions is found among the arguments of the public
choice economists. 2 2 Their image of the legislator is of a person who has
22. Public choice applies the same principles that economists use in analyzing
people's actions in the marketplace to people's actions in collective decisionmaking.
Economists who study behavior in the private marketplace assume that people are
motivated mainly by self-interest. Although most people base some of their actions on
their concern for others, the dominant motive of employers, employees, and consumers in
the marketplace is a concern for themselves. Public choice economists make the same
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1135
a lively regard for his own self-interest, and a considerably less lively regard for the public good. In this view, the task of the legislator is to be
free to vote for policies that can be framed in ways that will lead to voteattracting sound bites to be cited in the next reelection campaign. Conventional wisdom strongly holds that "being tough on crime" is one such
issue and being anything but tough on crime will be a weapon in the
hands of an opponent of the incumbent. Thus, driven by these calculations, the legislator will automatically vote for "severe sentences" and "law
23
and order" legislation.
It is not necessary to embrace all of the perspectives of public choice
economists to be skeptical about the likelihood of the legislature optimizing public good in highly politicized decisions. Nelson Polsby notes that
some congressional committees at certain points in their existence were
characterized by "a respite from the rigors of partisan combat ....
These
committees frequently took great pride in delivering to the floor proposals that had bipartisan approval and reflected bipartisan cooperation in
the framing of legislation." 24 However, he points out that at other periods of time, congressional decisionmaking has been driven by blatantly
political considerations designed to attract votes to the controlling
party.
25
Both of these perspectives raise concerns about the legislature as a
policymaking body for decisions likely to raise criticism at election time.
The solution that is sometimes used in these cases is the "independent
commission of experts" that is charged with arriving at a set of recommendations. The commission, by putting the weight of its expertise beassumption-that although people acting in the political marketplace have some concern
for others, the main motive of voters, politicians, lobbyists, and bureaucrats is self-interest.
Jane S. Shaw, Public Choice Theory, in The Concise Encyclopedia of Economics (David R.
Henderson ed., 2002), at http://www.econlib.org/library/Enc/PublicChoiceTheory.html
(on file with the Columbia Law Review).
23. This analysis is consistent with the voting evidence. Over the last decades there
has been a remarkable escalation in the duration of the criminal penalties imposed by
legislatures. Most researchers have now concluded that a crime control strategy of
increasing sentence severity is not an effective way of reducing crime rates and that such a
strategy imposes high costs on the society that implements it. See, e.g., Jane B. Sprott &
Anthony N. Doob, Imprisonment Rates in Canada: One Law, 10 Outcomes, in Penal
Reform in Overcrowded Times 232, 238 (Michael Tonry ed., 2001). Sprott and Doob
suggest,
If a government is serious about conserving scarce criminal justice resources and
using these resources in a manner that most effectively addresses legitimate
concerns about public security, it might focus first on mechanisms to reduce the
number of long sentences or the proportion of time that those serving long
sentences are in a prison setting.
Id. at 238.
24. Nelson W. Polsby, How Congress Evolves: Social Bases of Institutional Change
125 (2004). Polsby is a well-respected, long-term observer of the United States Congress
who is less driven by theory about what Congress does than by careful observation and
interview-based conclusions about why various congressional decisions were made.
25. See, e.g., id. at 126 (discussing the Gingrich years).
1136
COLUMBIA LAW REVIEW
[Vol. 105:1124
hind its recommendations, creates a citable reason for the legislature to
adopt those recommendations, These concerns reinforce the propriety
of significant policyrnaking delegation to the sentencing commission and
add to our earlier conclusion about the scope of that delegation: Legislatures should delegate those policy decisions that are easily subject to manipulation for vote-getting purposes, such as issues that have come to
have political baggage associated with them. (See cells IA & lB.)
B. Decision 2: Rule Articulation
The policy-articulation decision shares many of the qualities of the
policymaking decision. Thus, both judges and juries are easily excluded.
Juries do not have the necessary knowledge. Judges may have more
knowledge but still only a small portion of the available knowledge that
could bear on issues such as how the various purposes of sentencing can
and cannot be satisfied. Moreover, neither juries nor judges have access
to the kind of research or expert staff that could systematically provide
26
such information.
Even if the knowledge problem did not exist, both judges and juries
are properly excluded as decisionmakers because of the disparity problem: To rely on either is to allow offenders brought before different decisionmakers to be subject to different punishment rules. Individual judges
and juries have no mechanism by which they can impose their sentencing
rules upon others (even if they were to articulate them). And justice
holds abhorrent differences in punishment according to differences in
decisionmakers rather than differences in offenses or offenders. (See
cells 2D & 2E.)
The legislature also would seem to have difficulty fulfilling the policy
articulation role. Given the variety of factors that are relevant to the punishment decision, the task is a large and detailed one. The legislature can
hire experts to inform itself, but is it realistic to think that it has the time
or inclination to immerse itself in the details required for effective rule
drafting? Further, the legislature would be a poor body to keep up with
26. Giving a sentencingjudge wide discretion, as was once the practice, might seem to
finesse the rule articulation problem by allowing the judge to sort out decision rules in her
head for her own cases, thus avoiding the need for articulated rules. But the empirical
studies suggest that when a judge performs this function she is not in fact bringing to bear
all available information on what will best advance the stated policy, such as optimizing
general deterrence-no human could. See, e.g., Mandeep K. Dhami, Psychological
Models of Professional Decision Making, 14 Psychol. Sci. 175, 175-80 (2003). She is,
rather, looking to just a few pieces of information that she has divined or informally
learned from others. Id. This has the effect of reducing a complex determination that
depends upon mastering an enormous amount of modern scientific information to a more
manageable form. The information commonly ignored, however, is important, and its
exclusion undermines the efficacy of the resulting rules. Indeed, the single most attractive
aspect of relying on a single sentencingjudge for this rule articulation may be the ability to
obscure the fact that so little of the relevant information is in fact being taken into
account.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1137
the ever changing conditions that may require frequent adjustment of
the sentencing rules, either because crime conditions have changed, new
knowledge suggests a change in approach, experience in application
shows the articulated rules require clarification to those applying them,
or any number of other reasons. Once the societal value judgments have
been made, per the policymaking discussion in Decision 1, there seems
little advantage in having the legislature do the policy articulation. Legislative interests expressed in its policymaking decisions can be adequately
protected by the legislature retaining for itself the ability to object to an
articulation that it sees as inconsistent with its policy directives. (See cell
2A.)
The sentencing commission and parole board seem the better decisionmakers in this instance. Their members can have expertise and can
draw on the expertise of others at length. Both can investigate issues and
articulate rules in as much detail as they find necessary. Their rules
would guide the sentencing of more cases than those that appear before
a single sentencing judge, thereby providing the uniformity in application that equality and justice require. Both can monitor the application
of their rules and adjust them as new knowledge and changing conditions
dictate.
As between the two, which body is preferable? One might immediately prefer the sentencing commission on the already noted ground that
the parole board is seriously disadvantaged by having jurisdiction over
only those cases that have previously been determined to merit terms of
imprisonment. The sentencing commission's articulation of policy, in
contrast, can govern all manner of sentences-fines, community service,
ankle bracelets, home confinement, noncustodial drug treatment, and
more. This might end up being the determining factor in the analysis.
But the other distinguishing feature of a parole board-its practice of not
only articulating rules but also applying them-has special importance to
the present issue. There is reason to believe that such an additional role
might give the rulemaker information and experience important to making the articulated rules more clear and consistent in application. Knowing firsthand the ambiguities that arise and the difficulties in application
can only strengthen the rulemaker's ability to refine and improve. Perhaps the ultimate conclusion here is that the sentencing commission
must be preferred because the parole board's limited jurisdiction is disqualifying, but the sentencing commission should learn an important lesson from the comparison and should establish feedback channels from
those persons applying their rules. Some commissions are already directed to do this.27 (See cells 2B & 2C.)
27. See, e.g., 28 U.S.C. § 994(o) (2000) (requiring that the U.S. Sentencing
Commission "periodically ... review and revise, in consideration of comments and data
coming to its attention, the guidelines" and "consult with authorities on, and individual
and institutional representatives of, various aspects of the Federal criminal justice system"
and that various agencies "at least annually, submit to the Commission a written report
1138
COLUMBIA LAW REVIEW
[Vol. 105:1124
C. Decision 3: Factfinding
An obvious decisionmaker to be quickly excluded as factfinder is the
legislature. It may have the capacity to serve as a factfinder, as it does for
presidential impeachment, 28 but it cannot afford the time to do so for
any less important case. (See cell 3A.)
Sentencing commissions and parole boards may have similar administrative difficulties. It is true that parole boards often determine the facts
of a case, upon which they will then base application of their parole
guidelines. Presumably, sentencing commissions could do something
similar. One could imagine a world in which sentencing commissions
held sentencing hearings based on the model of parole boards. But the
reality is that parole boards do not make an independent factual determination. Instead, their factfinding piggybacks on the factfinding of juries,
judges, and the court's probation officers. They typically use the trial record and the presentence report-rather than the testimony of witnesses
and the examination of forensic evidence-to reconstruct the events constituting the offense and other relevant facts. The enormous costs of
factfinding probably mean that it cannot realistically be done anywhere
other than at trial and post-trial sentencing hearings before judge orjury.
(See cells 3B & 3C.)
Should the factfinder at such proceedings be judge or jury? Each
has heard all of the testimony and could do the factfinding needed for
sentencing. Given the extensive social science literature on 'jury decisionmaking," one might expect a clear answer to the question, but in fact
it is not so easily addressed. First, there are few studies that directly compare judges and juries; second, there are few studies that expressly address reliability in factfinding-that is, the accuracy of the re-creation of
past events, as distinct from other aspects of decisionmaking.
1. Jury (Versus Judge) Decisionmaking Literature. - Judges and juries
can be compared in field studies using real cases (where one makes the
real decision and the other makes a hypothetical decision) or in laboratory studies (in which both make hypothetical decisions). The few extant
field studies on verdicts show that verdicts sometimes differ between
judge and jury. In the landmark study by Kalven and Zeisel, U.S. district
court judges were asked whether they agreed with the jury's verdict in
recent criminal trials over which they had presided. 29 There was agreement in 75% of the trials. 30 In another 17% of the trials, the judge would
commenting on the operation of the Commission's guidelines, suggesting changes in the
guidelines that appear to be warranted, and otherwise assessing the Commission's work");
42 Pa. Cons. Stat. Ann. § 2155(a) (Purdon 2002) (requiring that prior to the adoption of
any proposed sentencing guidelines, the Pennsylvania Commission on Sentencing must
hold public hearings that afford an opportunity to testify to a number of organizations,
including those representing persons involved in carrying out the guidelines).
28. See U.S. Const. art. I, § 3, cl.6.
29. Harry Kalven & Hans Zeisel, The American Jury 488 tbl.139 (1966).
30. Id. at 56 tbl.11.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1139
have convicted whereas the jury acquitted; in 2%, the jury convicted
whereas the judge would have acquitted; and in 6%, there was a hung
jury (and the judge would have convicted in most of these cases).'"
Although interesting, these results are not determinative as to who
would be the better factfinder. First, for present purposes-where we are
considering factfinding in regards to sentencing only-differences in verdicts do not necessarily demonstrate differences in the assessment of the
underlying facts. Jurors recognize that their verdicts form the basis for
criminal punishment and thus might have reasons for a "not guilty" verdict even when the facts could be assessed otherwise; for example, they
might not want the defendant punished, they might believe that the law is
unfair or silly, or they might believe that the defendant already has been
punished enough.3 2 In addition, jurors may have determined the facts
correctly, but misapplied the law.33 Second, in these studies of real cases,
when judge and jury differ, it cannot be assumed thatjudges are the ones
who "got it right." We cannot know what the truth is in these cases, so we
cannot know who is the better factfinder. s 4 Thus, although the studies
do imply that there might indeed be a difference in the factfinding ability
ofjudges and juries, this conclusion remains tentative because of the first
of the reasons noted above.
31. Id. Later studies also reveal high judge-jury consistency in conviction rates and
suggest factors that might be relevant in explaining when judgments might be different.
See Theodore Eisenberg et al., Judge-Jury Agreement in Criminal Cases: A Partial
Replication of Kalven & Zeisel's The AmericanJury, 2 J. Empirical Legal Stud. 171, 179-93
(2005); Larry Heuer & Steven Penrod, Trial Complexity: A Field Investigation of Its
Meaning and Its Effects, 18 Law & Hum. Behav. 29, 48 tbl.12 (1994).
32. For a good evaluation of the merits and problems of the Kalven and Zeisel study,
see Lawrence S. Wrightsman et al., Psychology and the Legal System 366-77 (5th ed.
2002).
33. This point is highly consequential for verdict determination and more or less
relevant here depending on what is meant by "factfinding" for sentencing. The problem is
not with jurors determining whether, for example, the defendant went to the victim's
home, had a gun, and had intent to kill; rather, the problem is with jurors' ability to
consider those kinds of facts and evaluate whether they meet the legal description of the
elements of specific crimes. Vicki L. Smith has shown that mock jurors are likely to render
verdicts in accord with their lay prototypical view of crimes rather than in accord with the
legal definitions of crimes provided by the experimenter. See Vicki L. Smith, Prototypes in
the Courtroom: Lay Representations of Legal Concepts, 61 J. Personality & Soc. Psychol.
857 (1991). For a discussion of ways in which juries and judges may reach different
decisions even if they agree on the facts, see Robert J. MacCoun, Epistemological
Dilemmas in the Assessment of Legal Decision Making, 23 Law & Hum. Behav. 723,
726-28 (1999).
34. Third, field studies like that by Kalven and Zeisel are not entirely "naturalistic"
studies in that only one group-judge or jury-can actually be making the real
determination. In the Kalven and Zeisel study, juries rendered verdicts and later the
judges who had presided over the cases were surveyed as to how they would have ruled in
the case. Kalven & Zeisel, supra note 29 at 45. Thus, the judges already knew what the
juries had decided. In addition, only about one-sixth of the judges contacted answered the
survey. Id. at 35-36.
1140
COLUMBIA LAW REVIEW
[Vol. 105:1124
In contrast to the small number of field studies, there are a myriad of
laboratory studies examining the performance of "mockjurors. ' 35 These
studies find a variety of weaknesses injury decisionmaking. For example,
jurors' decisions may be influenced by factors other than the facts of the
case, such as pretrial publicity, the physical appearance of a witness or
defendant, inadmissible evidence, or hindsight bias. Jurors also may confuse arguments with evidence. Such factors may affectjurors' factfinding
or other decision processes. If subjects' verdicts in these mock cases were
solely the result of factfinding, such data would demonstrate that jurors
are far from perfect factfinders. However, again, these data cannot be
used to rule out juries in favor of judges because they do not tell us
whether judges are equally flawed factfinders, for either similar or different reasons. Mock case studies are rarely run using actual judges; judges
are not easily corralled into participating in laboratory research. What
limited evidence exists tends to suggest that judges, like jurors, may be
influenced by extralegal factors, although to a lesser extent. 36 Further,
while in these laboratory studies we know the "right answer" as to the
facts because the researchers have created those facts, the studies may not
capture the true complexities of determining past events as they occur in
the messiness of the real world.
2. OtherRelevant Social Science Literature. - Although we know of no
research specifically comparing judges and juries on factfinding per se,
the psychology literature provides other research relevant to who is likely
to be better at this task. The most important contrasts between judge and
july for these purposes are that (i) judges have both training and experience (that is, they commonly are lawyers and have ruled on many cases)
whereas juries have neither; (ii) the average judge and the average juror
commonly differ in qualities relevant to decisionmaking (for example,
intelligence, general reasoning skills); (iii) judges are individuals whereas
juries are groups; and (iv) judges are elected or appointed whereas juries
are "drafted" from the community somewhat at random.
35. Studies with "mock jurors" are most often run with college students as
participants, although there is an increasing use of a broader group of 'jury-eligible"
adults. An overall analysis of studies comparing student and nonstudent samples
concluded that there are few differences between the groups, and, if anything, student
jurors are more likely to find for the defendant in criminal trials. Brian H. Bornstein, The
Ecological Validity of Jury Simulations: Is the Jury Still Out?, 23 Law & Hum. Behav. 75,
77-80 (1999).
36. See, e.g., Chris Guthrie et al., Inside the Judicial Mind, 86 Cornell L. Rev. 777, 784
(2001) (showing that judges demonstrate common cognitive heuristics and biases such as
anchoring, framing effects, and hindsight bias, although sometimes less so than jurors);
Reid Hastie & W. Kip Viscusi, WhatJuries Can't Do Well: TheJury's Performance as a Risk
Manager, 40 Ariz. L. Rev. 901, 906-07 (1998) (showing that judges also show hindsight
bias, although to a lesser extent than juries); Andrew J. Wistrich et al., Can Judges Ignore
Inadmissible Information? The Difficulty of Deliberately Disregarding, 153 U. Pa. L. Rev.
(forthcoming 2005) (showing that judges are influenced by inadmissible evidence). One
would expect that judges would not make the mistake jurors make of confusing evidence
and argument.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1141
Both juror and judge factfinding can be best characterized as story
construction: People try to fit the evidence into a plausible story of what
happened. 37 Evidence that does not fit a coherent story may be disbelieved or devalued (that is, found not to be a "fact"). 38 Research on juror
text comprehension and decisionmaking (as well as research on reasoning generally) demonstrates that people do not hold a mass of facts in
memory until a final instant at which they then form their opinion of
what happened. Rather, facts are integrated into stories or opinions as
information is revealed. Jurors and judges may have pretrial pro-plaintiff
or pro-defendant biases. 39 Such biases can lead people to interpret new
facts in accordance with whichever verdict they currently favor. Striving
to interpret facts coherently can occur not only when the juror or judge
comes in with an initial opinion of the case, but also as she starts to favor
40
one side over the other as the trial unfolds.
On the other hand, some facts may not be able to be forced into the
current story. Thus, when ajuror's constructed story is sufficiently called
into question by later incontrovertible evidence, she may revise her view
of the entire story. However, how and whether a person revises a constructed story in light of new facts may depend on the sophistication of
41
that person's reasoning processes.
37. Nancy Pennington & Reid Hastie, Evidence Evaluation in Complex Decision
Making, 51 J. Personality & Soc. Psychol. 242, 242 (1986); Nancy Pennington & Reid
Hastie, Reasoning in Explanation-Based Decision Making, 49 Cognition 123, 123-24
(1993); Nancy Pennington & Reid Hastie, The Story Model for Juror Decision Making, in
Inside the Juror: The Psychology of Juror Decision Making 192, 192-94 (Reid Hastie ed.,
1993).
38. Paul Thagard, Explanatory Coherence, 12 Behav. & Brain Sci. 435, 449-50
(1989).
39. In fact, judges may have greater initial belief in the guilt of the defendant due to
experience and/or faith in the legal system. Note that second-time jurors are also more
likely to convict than first-time jurors. Melissa J. Himelein et al., Effects of Prior Juror
Experience on Jury Sentencing, 9 Behav. Sci. & L. 97, 97 (1991). For a review of the effects
of prior jury service, see Ronald C. Dillehay & Michael T. Nietzel, Prior Jury Service, in A
Handbook of Jury Research 11-1 (Walter F. Abbott & John Batt eds., 1999).
40. See Kurt A. Carlson &J. Edward Russo, Biased Interpretation of Evidence by Mock
Jurors, 7 J. Experimental Psychol.: Applied 91, 92 (2001) (showing how evidence
accumulates for the side one is already leaning toward); Keith J. Holyoak & Dan Simon,
Bidirectional Reasoning in Decision Making by Constraint Satisfaction, 128 J.
Experimental Psychol.: Gen. 3, 8-10 (1999) (detailing how shifts in some beliefs are made
to be consistent with shifts in others); Dan Simon et al., Construction of Preferences by
Constraint Satisfaction, 15 Psychol. Sci. 331, 335 (2004) (arguing that preferences are not
fixed); Dan Simon et al., The Emergence of Coherence over the Course of Decision
Making, 27J. Experimental Psychol.: Learning, Memory, & Cognition 1250, 1257 (2001)
(also detailing how shifts in some beliefs are made to be consistent with shifts in others);
Dan Simon & Keith J. Holyoak, Structural Dynamics of Cognition: From Consistency
Theories to Constraint Satisfaction, 6 Personality & Soc. Psychol. Rev. 283, 286-89 (2002)
(reviewing experimental literature on coherence shifts and its implications for legal
decisionmaking).
41. Deanna Kuhn et al., How Well Do Jurors Reason? Competence Dimensions of
Individual Variation in a Juror Reasoning Task, 5 Psychol. Sci. 289, 295 (1994).
1142
COLUMBIA LAW REVIEW
a. Training and Experience. -
[Vol. 105:1124
One might assume that the training
and experience of judges makes them obviously better than juries at
factfinding. Yet for the task of factfinding it is not apparent to what extent experience helps, and it is quite possible that experience could
(sometimes) hurt. Judges hear case after case but get no reliable feedback as to whether their (or ajury's) determinations of fact are accurate.
Thus, while their experience may build, they have no mechanism by
which to test the reliability of the factfinding rules or procedures that
they use. They may be regularly getting it wrong, but never have occasion
to learn of their errors. It is true that judges, with their experience from
hearing many cases, are likely to be more consistent in factfinding over
time than jurors (that is, in finding similar facts in similar situations), but
without feedback there is no reason to believe that those consistent
find42
ings are closer to the truth than the one-time-only jury finding.
In fact, consistency can be the downside of experience. Having seen
many trials, judges may develop schemas for certain "types" of trials.
These schemas may create expectations. For example, a judge may conclude that after a defendant kills the victim, he usually tries to make up
an alibi, ditch the gun, etc. Judges might be more likely than jurors to
notice when some important facts are missing in a particular case-for
example, when some action by the defendant, necessary for him to have
committed the crime, has not been proven; however, judges might also
be more likely than jurors to fill in missing details with schema-consistent
knowledge. Thus, judges might misconstrue the facts of one case based
on similarity to previous cases when such an inference is not warranted.
Typically, overreliance on schemas is not a characteristic of experts;
in fact, experts are known for being able to see subtle distinctions between instances. 43 Nonetheless, again, due to the lack of feedback, it
would be a mistake to assume that the experience ofjudges leads them to
expertise. Without feedback, judges can fall prey to reliance on schemas
and thereby fail to account for the individuating facts in a particular case.
b. Individual Differences Between Judges and Juries. -
Another impor-
tant difference between judges and jurors may be that judges (on average) score differently from jurors (on average) on individual difference
42. Consider, for example, the expertise of a golf player. After much practice without
feedback (i.e., without hearing/feeling the smack of the ball and without seeing where it
lands), a player might develop a consistent swing but would be unlikely to ever get close to
a target.
43. The good news/bad news of expertise is described well in Peter A. Frensch &
Robert J. Sternberg, Skill-Related Differences in Game Playing, in Complex Problem
Solving: Principles and Mechanisms 343 (Robert J. Sternberg & Peter A. Frensch eds.,
1991). In the game of bridge, for example, experts have better memory for hands played
than novices-even though they have played more (lifetime) hands. However, when
experts and novice bridge players are asked to play a bridge-like game with minor
variations, expert performance is much more impaired than novice performance. Id. at
360-62.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1143
measures such as Need for Cognition, 4 4 Need for Cognitive Closure, 45 or
general reasoning competence. 46 Given their education, and the fact
that most of these measures correlate with education, IQ or both, it is
likely that judges are better at constructing complete stories than randomly selected jurors. However, data suggest that the act of deliberating
may catuse jurors to become more sophisticated reasoners than they were
before deliberating. 47 In addition, whereasjudges must rely only on their
own cognitive skills, jurors work in groups and gain the benefit of the
cognitive skills of others in the group.
c. Groups Versus Individuals. - Just as one might have believed (incorrectly) that experienced people are hands down better factfinders
than novices (thus creating a preference for judges), one might similarly
believe (incorrectly) that groups are hands down better factfinders than
individuals (thus creating a preference for juries). However, the literature on group versus individual decisionmaking is mixed; groups do not
48
perform better than individuals as often as one might hope or expect.
44. The Need for Cognition is a measure of people's enjoyment of and motivation for
effortful thinking. Mock jurors with high Need for Cognition are perceived as
participating more actively injury discussion and are more resistant to persuasion. Donna
Shestowsky & Leonard M. Horowitz, How the Need for Cognition Scale Predicts Behavior
in Mock Jury Deliberations, 28 Law & Hum. Behav. 305, 317 (2004). For a review of the
Need for Cognition measure, see generally John T. Cacioppo et al., Dispositional
Differences in Cognitive Motivation: The Life and Times of Individuals Varying in Need
for Cognition, 119 Psychol. Bull. 197 (1996).
45. The Need for Cognitive Closure is a measure of an individual's desire for
predictability, a preference for order and structure, discomfort with ambiguity,
decisiveness, and close-mindedness. It can be dispositional or situational. Donna M.
Webster & Arie W. Kruglanski, Individual Differences in Need for Cognitive Closure, 67J.
of Personality & Soc. Psychol. 1049, 1050 (1994).
46. Deanna Kuhn measures a person's facility with arguments and rates her on a
continuum from "satisficing" (willing to ignore information inconsistent with a somewhat
plausible story) to "theory-evidence coordination" (willing to construct multiple stories).
Kuhn et al., supra note 41, at 289.
47. See Monica L. McCoy et al., The Effect ofJury Deliberations on Jurors' Reasoning
Skills, 23 Law & Hum. Behav. 557, 561 (1999) (demonstrating that post-deliberation mock
jurors are more likely to reason at a higher level than pre-deliberation mock jurors).
48. For a recent review of the group decisionmaking literature, see Norbert L. Kerr &
R. Scott Tindale, Group Performance and Decision Making, 55 Ann. Rev. Psychol. 623
(2004). Of late, group decisionmaking has gotten a bad name: Cries of "groupthink"
permeated the media in response to the failure to find weapons of mass destruction in Iraq
as "everyone" believed would happen. See, e.g., Dana Priest & Dafna Linzer, Spy Panel
Condemns Iraq Prewar Intelligence: '02 Estimate Faulted, CIA 'Group-Think' Criticized,
Wash. Post, July 11, 2004, at Al. The term "groupthink" was introduced by IrvingJanis to
describe group decisionmaking in which the goal of maintaining group cohesiveness
develops more importance than the goal of making a good decision (and, for example,
realistically considering all the facts, bringing up contradictory evidence, etc.). See Irving
L. Janis, Victims of Groupthink, 8-13 (1972). Groupthink is most likely to occur in
cohesive isolated groups with strong leaders under high stress. These conditions are not
typical characteristics ofjuries. For a good description of groupthink, see Elliot Aronson et
al., Social Psychology 290-301 (5th ed. 2005).
1144
COLUMBIA LAW REVIEW
[Vol. 105:1124
For our purposes, two factors are relevant to who is likely to be the
better decisionmaker: (1) the features of the problem and (2) the composition of the group. Although not uniformly so, the literature on
group decisionmaking suggests that groups are better than individuals at
problems that involve a "right answer" or an "insight answer" (for example, trivia questions or brain teasers) rather than a probabilistic answer
(for example, judging which of two stocks is likely to do better) because,
with the former, if one person comes up with the right answer, she can
usually persuade the others through justification. 49 Groups may also
have an advantage over individuals in decisions involving judgments (or
evaluation) rather than decisions involving selecting between two alterna50
tives because groups are better at taking multiple factors into account.
How should factfinding for sentencing purposes be characterized?
Superficially, it seems more like a deciding-between-alternatives task: The
point is to decide, for example, whether the defendant did or did not
perform certain acts and did or did not have a particular state of mind.
Yet, again, such decisions are not made in isolation. Factfinding involves
creating a coherent story that takes multiple factors into account. And, as
in insight problems, one juror's insight (for example, recognizing an obscure inconsistency in testimony) could turn the group judgment around
51
if it causes previously incoherent or contradictory evidence to cohere.
The composition of a group also is implicated in how well a group
will perform. Groups in which all information is shared (that is, in which
all members have the same information at their disposal) perform better
than groups that do not. 52 Jurors all see and hear the same information
during the trial-although due to attention lapses and memory differences, jurors may vary on what information they can access. In addition,
jurors may have unshared preexisting knowledge that could be relevant
to the trial. Jurors may differ in how likely they are to bring unshared
knowledge to the group (factors such as status and expertise will affect
49. In the former decisions, truth wins. Patrick R. Laughlin & Alan L. Ellis,
Demonstrability and Social Combination Processes on Mathematical Intellective Tasks, 22
J. Experimental Soc. Psychol. 177, 179 (1986).
50. See Daniel Gigone & Reid Hastie, The Impact of Information on Small Group
Choice, 72J. Personality and Soc. Psychol. 132, 139 (1997).
51. Is that likely to happen? In the Kalven and Zeisel study, where an initial majority
of jurors favored a particular verdict, the jury's final decision was the same as the initial
majority decision 97% of the time. Of course, in those cases, we don't know which
decision was right. See Kalven & Zeisel, supra note 29, at 488 tbl.139. In a laboratory study
involving three-person groups, there were few times (5%) in which an initial group
majority was overturned for the final decision. However, most of the times when the initial
majority was overturned, the switch was from an incorrect to a correct answer. Gigone &
Hastie, supra note 50, at 138.
52. For a brief readable review of the group preference for discussing shared
information, see Gwen M. Wittenbaum & Ernest S. Park, The Collective Preference for
Shared Information, 10 Current Directions in Psychol. Sci. 70 (2001). For a recent update
on literature discussing the bias toward shared information, see Kerr & Tindale, supra note
48, at 636-38.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1145
such contributions), and the group might not give appropriate weight to
such knowledge. To the extent that all information is shared by the
group, however, the group may have a larger information base than an
individual judge, and therefore may be better at factfinding.
d. Selection and Perception. - An issue we have not addressed yet is
bias: Between judge and jury, which is more likely to be biased, and
which is more likely to be perceived as biased by the public? Ideally, we
like to think ofjudges as being unbiased; however, we know that our ideal
is not reality. More importantly, even judges who try to be unbiased may
not succeed. 53 For example, studies have shown that Democratic and Republican state supreme court justices differ (modestly) in their decisions,
with Democrats leaning toward the defense in criminal cases. 5 4 This bias
acts across different groups; we can imagine that individual judges might
also show unintentional biases towards different types of cases or defendants. 55 Of course,jurors individually and juries in the aggregate may also
be biased. However, in these complex decisionmaking situations, groups
56
may temper individual biases.
Regardless of the perception that juries were biased in some highly
visible cases (for example, the first Rodney King trial or the O.J. Simpson
criminal trial), citizens tend to believe that having juries render verdicts
offers more procedural fairness than having judges render verdicts. 57 We
assume that the same perception would hold for factfinding as a basis for
sentencing decisions.
Our conclusion must be a modest one, for the evidence paints a
mixed picture. Bothjudges and juries have advantages and disadvantages
as reliable re-creators of past events. Given our current limited data, one
might conclude that juries, on balance, may have a slight advantage over
judges as factfinders, not because of an actual superiority in reliable
factfinding, but because they are perceived as being superior, and that
53. This problem, of course, does not only apply to judges. Often, when people try to
be unbiased they cannot succeed. Timothy D. Wilson & Nancy Brekke, Mental
Contamination and Mental Correction: Unwanted Influences on Judgments and
Evaluations, 116 Psychol. Bull. 117 (1994); updated in Timothy D. Wilson et al., Mental
Contamination and the Debiasing Problem, in Heuristics and Biases: The Psychology of
Intuitive Judgment 185 (Thomas Gilovich et al. eds., 2002).
54. Daniel W. Shuman & Anthony Champagne, Removing the People from the Legal
Process: The Rhetoric and Research on Judicial Selection andJuries, 3 Psychol. Pub. Pol'y
& L. 242, 245 (1997).
55. See Hogarth, supra note 20, at 22.
56. The data are mixed as to whether groups show more or less "bias" than individuals
in various judgment and decisionmaking tasks. Norbert L. Kerr et al., Bias in Judgment:
Comparing Individuals and Groups, 103 Psychol. Rev. 687, 692 tbl.1 (1996). However,
groups tend to make more consistent and reliable judgments than individuals, suggesting
less idiosyncrasy in group information processing. Verlin B. Hinsz et al., The Emerging
Conceptualization of Groups as Information Processors, 121 Psychol. Bull. 43, 50 (1997).
57. Robert J. MacCoun & Tom R. Tyler, The Basis of Citizens' Perceptions of the
Criminal Jury: Procedural Fairness, Accuracy, and Efficiency, 12 Law & Hum. Behav. 333
(1988).
1146
COLUMBIA LAW REVIEW
[Vol. 105:1124
perception can have important practical benefits for the effective operation of a sentencing system. 58 (See cells 3D & 3E.)
D. Decision 4: Judgment-Making
Many of the same observations regarding who is better at factfinding
apply as well to judgment-making. The legislature cannot adjudicate individual cases. As before, the parole commission has the usual problem
of being limited in jurisdiction to cases involving individuals that previously have been sentenced to prison. The enormous costs of the presentation of evidence necessarily require that it be done either at trial or
post-trial hearings; the parole board and sentencing commission can at
best piggyback on these prior determinations. Indeed, in light of the nature of this particular decision-judgment-making rather than factfinding-it seems even more important that the decisionmaker be one who
has observed the defendant and victim and who has the kind of familiarity that comes with hearing the live witnesses. 59 (See cells 4A, 4B & 4C.)
We are therefore left with the same two choices as in determining
the best factfinder. Should judge or jury be preferred as a judgmentmaker? The answer to this question may depend upon the dominant
purpose or purposes to which the policymaker has committed the system
in Decision 1. For example, if shared community intuitions ofjustice are
a central determinant of liability, 60 the jury will have a substantial advantage because it can more reliably articulate those intuitions than an individualjudge. On the other hand, if the dominant purpose is the incapacitation of dangerous offenders-for which the central criteria is not the
offender's blameworthiness but a clinical assessment of his dangerousness-or if the dominant purpose is the general deterrence of potential
offenders-for which the central criteria is again factors other than
blameworthiness-then judges may be better able to elicit and understand those expert opinions upon which the judgment will be based.
However, as noted previously, groups tend to be better at decisions involving judgment or evaluation because groups are better at taking multi61
ple factors into account.
58. Tom R. Tyler, Why People Obey the Law 60 (1990) (documenting importance of
reputation for procedural fairness for law's legitimacy and hence influence); Paul H.
Robinson & John M. Darley, The Utility of Desert, 91 Nw. U. L. Rev. 453, 471-77 (1997)
[hereinafter Robinson & Darley, Utility of Desert] (reviewing ways in which criminal law's
moral credibility with community can affect its crime control power).
59. Sometimes small judgment differences can make large liability differences. This is
true, for example, if the system seeks to reflect community intuitions of justice. See
generally Paul H. Robinson & John M. Darley, Justice, Liability & Blame: Community
Views and the Criminal Law (1995) [hereinafter Robinson & Darley, Justice, Liability &
Blame] (reporting eighteen studies on different aspects of laypersons' intuitions of
justice).
60. See, e.g., Robinson & Darley, Utility of Desert, supra note 58, at 471-77.
61. See supra note 50 and accompanying text.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1147
Where does this leave us with regard to preferring judge or jury?
Juries seem preferable as judgment-makers, but the extent of that preference may vary with sentencing purpose. Different jurisdictions may rely
on a different mix of sentencing purposes. But at least this can be said:
While there may be some who advocate that blameworthiness should be
the only determinant of criminal punishment, there are few who would
advocate that blameworthiness should have no important role in determining liability-or at least this is the conclusion one would come to
upon examining American criminal justice systems as they exist today. 62
Even codes based upon the Model Penal Code, which in the 1960s enthusiastically supported a shift toward incapacitation and deterrence programs, are careful not to deviate too far from blameworthiness criteria,
and indeed, explicitly incorporate liability rules that look to lay intuitions
63
of justice.
The preference for juries over judges must be moderated by a concern for consistency in application. There is an unfairness in having different juries impose different normative standards in their judging.
While it is true that, in the context of judging blameworthiness, there is
reason to believe that there is a fair amount of agreement among lay
intuitions of justice-even across demographics 64 -it is also true that
without access to similar judgments made by other juries in similar cases,
there is at least a risk that juries may differ in their judgment-making in
borderline cases. Judges have more consistent judgment, at least among
their own cases, but with few mechanisms for judges to systematically
learn what judgments other judges make in similar situations, their advantage may be easy to exaggerate. The best response to the consistency
problem is for the articulated sentencing rules in Decision 2 to provide
examples of situations that do and do not meet the standard-or to at
least direct the decisionmakers' attention to the relevant factors 6 5-examples that all Decision 4 decisionmakers could use to improve the consistency of their judgments. Judges normally might have a slight advantage over juries in consistency in judgment, but that advantage is easily
washed out if, as is preferable, the articulated sentencing rules provide
illustrations.
62. See Robinson, Criminal Law, supra note 21, at 17 ("[N]o purely utilitarian system
for the distribution of liability and punishment has ever been adopted.").
63. See Paul H. Robinson, Why Does the Criminal Law Care What the Layperson
Thinks Is Just? Coercive Versus Normative Crime Control, 86 Va. L. Rev. 1839, 1841-57
(2000) (citing examples of deference to lay intuitions of justice in modern American
criminal law).
64. See Robinson & Darley, Justice, Liability & Blame, supra note 59, at 225-28.
65. Some examples are available in Model Penal Code § 5.01(2) (a)-(g) (1985)
(giving illustrations of what constitutes a "substantial step" for purposes of the attempt
offense). For an example of factors to be considered, see 1 Final Report of the Illinois
Criminal Code Rewrite and Reform Commission §§ 507(3), 508(4), 509(4) (2003) (listing
factors to be considered in determining a variety of elements of the defenses of duress,
ignorance of law, and reliance upon official misstatement of law, respectively).
1148
COLUMBIA LAW REVIEW
[Vol. 105:1124
Perhaps most important in allocating judgment-making is the matter
of public perception, which, recall, was determinative with regard to
factfinding. Here the case would seem to be stronger. Jurors really do
have an advantage in making the normative judgments that the public so
closely associates with doing justice. It would seem to follow that here,
even more than with factfinding, the system needs the public confidence
in judgment-making that the jury can best give. (See cells 4D & 4E.)
E. Decision 5: Determining Punishment Amount
Applying the factual and normative findings of Decisions 3 and 4 to
the sentencing rules of Decision 2 in order to determine an amount of
punishment requires a good deal of training and expertise: mastering
the meaning of the Decision 2 rules, and learning to consistently and
accurately apply those rules. That would seem to exclude the jury as the
decisionmaker here. It would also require participation in individual
cases, which obviously would exclude the legislature. (See cells 5A & 5D.)
It would be possible for parole boards to perform the task, and traditionally they have done so, but they have the usual disqualifier of having
jurisdiction over only imprisonment cases. It would be possible for sentencing commissions to similarly perform the task, although traditionally
they have not done so. (See cells 5B & 5C.)
How would the individual sentencing judge as this decisionmaker
compare to the sentencing commission? To the extent that the task is
clerical, there is greater efficiency in having it performed by the individual judge. Why have a group do what a single person can do as effectively? To the extent that the task requires the exercise of judgment, a
small group may have the advantage, including greater uniformity in application and access to a wider variety of views in decisionmaking. Which
of these two applies probably depends upon the level of detail of the
sentencing rules promulgated at the policy articulation stage (Decision
2). The more detailed the rules, the more efficient it is to rely upon a
single sentencing judge to implement them. The less detailed the rules,
the greater the need for a sentencing commission to apply them in the
individual case. Just as many parole boards are not likely to hear all cases
themselves, but rather supervise application of the case findings to the
rules through supervised hearing officers, the ideal system may be one in
which the sentencing commission supervises "hearing officers." But
there seem to be good efficiency reasons to use individual sentencing
judges in the role. This would allow the commission to exercise that
amount of direct supervision required yet maximize the efficiency of the
individual decisionmaker where direct supervision is not required. That
is, the sentencing commission could determine the amount of oversight
required based upon the level of detail of the decision rules, which may
change over time and may change with case type. The conclusion, then,
is that the best decisionmaker may be the individual sentencing judge,
but subject to oversight by the sentencing commission to ensure that the
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1149
commission's decision rules are being followed where appropriate. (See
cells 5B & 5E.)
F. Decision 6: Determining Punishment Method
The analysis of who might best formulate a specific sentence will
track much of what has been said above regarding who might best calculate the punishment amount. The method-of-punishment decision requires a good deal of training and expertise, which would seem to exclude the jury. It also requires involvement in each case, which as a
practical matter excludes the legislature. As above, the parole board, as a
trained, small-group decisionmaker, is disadvantaged in relation to the
sentencing commission, which shares these same attributes but which, because of its earlier placement in the process, can influence all forms of
punishment, not just prison terms. (See cells 6A, 6C & 6D.)
That brings the punishment method choice back to the same finale
as with punishment amount. Do the sentencing commission and the individual sentencing judge have different comparative advantages with regard to this decision than they did with the previous decision? As noted
in Part I, the decision as to the proper sentencing method or combination of methods is highly complex and requires a good deal of information, both about (1) alternative punishment methods and their effect and
(2) the details of the people and existing circumstances. As to (1), the
decisionmaker must know, for example, what kind of sanctions work
under what conditions to achieve what purpose. This seems like the kind
of expertise that one could master only by spending a career keeping up
with the scientific literature (and producing some share of the literature
to investigate important questions not already answered by academics).
As to (2), on the other hand, the decision requires an intimate knowledge of the facts of the case at hand. An enormous number of factors
66
may affect the choice of sanctioning methods.
One might conclude that only the sentencing commission could provide the first kind of information-research and expertise-and that only
the sentencing judge could provide the second-details about the situation at hand. This may suggest a decisional process similar to that for
Decision 5-a sentence determination by an individual sentencing judge,
under the supervision of the sentencing commission-even though we
66. For instance, what is the nature of the offender's relationship with family
members? With other persons or with institutions that might participate in his future?
What treatment programs currently have space for new clients? What is the offender's
likely amenability to the treatment programs available? What are the offender's
employment history and employment prospects? What is his criminal history? What kinds
of crimes, how long ago, and under what circumstances? What past punishment has been
imposed upon him and to what effect? What predictions can be made of his likelihood of
future criminality? What visibility has this offense had in the community? What effect
would various sentence options have on the community? What is the current rate of
commission locally of the offense at issue or of offenses closely related to it? Is the rate
increasing or decreasing?
1150
COLUMBIA LAW REVIEW
[Vol. 105:1124
have reached this conclusion through a different path. In this instance,
the nature of the relationship between the two might be somewhat different. In Decision 5's determination of punishment amount, the important need for uniformity in application requires that the sentencing commission have oversight authority; judges best serve as implementers of the
commission's single will. But in the context of determining punishment
method, uniformity in application is no longer a primary concern. Every
case inevitably will be different from every other-different available
methods of punishment depending on location and different needs for
sanctioning methods depending on the community and individual situation. More importantly, the demand of justice that meaningfully similar
offenders be treated the same is a demand that concerns amount but not
method. As long as the same total amount of punishment is imposed on
similar offenders, there can be no unjustified disparity. 67 In other words,
as long as the proper amount of punishment is imposed, it may be imposed in any way that most effectively advances other legitimate objectives, such as the goal of reducing future crime, without endangering justice or fairness. This argues in favor of giving greater freedom of action
to individual judges and using sentencing commissions only in the role of
providing useful information and support. (See cells 6B & 6E.)
These analyses of the best decisionmaker for each of the six decisions might be summarized this way:
67. Such freedom in fashioning the actual sanctioning method may require a system
for adjusting the punishment "credit" of each method according to its felt punishment
bite; such punishment ratio schemes are feasible. For a discussion of such a system, see
Paul H. Robinson & Michael T. Cahill, Law Without Justice: How and Why the Criminal
Law Deliberately Sacrifices Desert (forthcoming 2005) (on file with the Columbia Law
Review).
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
TABLE:
1151
RELATIVE ADVANTAGE OF ALTERNATIVE DECISIONMAKERS FOR
EACH ASPECT OF SENTENCING DECISION
Decisionmaker:
Decision:
A. Legislature
B. Sentencing
Commission
C. Parole Board
D. Jury
E. Judge
1. Policymaking
delegation
possible but
problem of prison
sentence
limitation
problems of
disparity,
availability of
resources,
competency
problems of
disparity,
availability of
resources,
competency
2. Rule
Articulation
problem of prison
sentence
limitation
problems of
disparity,
availability of
resources,
competency
problems of
disparity,
availability of
resources,
competency
3. Factfinding
not practical
less practical
(than judge or
jury)
less practical
(than judge or
jury)
feasible alternative
to jury
4. Judgmentmaking
not practical
less practical
(than judge or
jury)
less practical
(than judge or
'jury)
unlikely but
conceivable
alternative to jury:
if blameworthiness
judgments not
required
5. Determining
Punishment
Amount
not practical
supervision
authority: oversee
exercise of
authority by
sentencing judge
less suited (than
S.C.) because of
prison sentence
limitation
problem of lack of
training
6. Determining
Punishment
Method
not practical
support role:
provide
information and
support useful to
judges
less suited (than
S.C.) because of
prison sentence
limitation
problem of lack of
training and
competency
-th",
IV. IMPLICATIONS FOR SENTENCING REFORMS
The standard practice until a few decades ago was to give judges
broad sentencing discretion. Part III's discussion articulates the many
specific difficulties with that traditional approach. Judges may be well
suited to apply articulated sentencing rules to the facts and judgments
made by others in order to determine the amount of punishment to be
imposed and to determine the best methods of that punishment, but
judges are inferior to other decisionmakers in performing the other aspects of the sentencing decision. In part because of a growing recognition of these difficulties with unrestrained judicial decisionmaking, the
past several decades have seen a variety of sentencing reforms. Some are
consistent with the conclusions reached in Part III, others are not.
1152
COLUMBIA LAW REVIEW
[Vol. 105:1124
A. Mandatory Minimum Sentences
Legislatures have increasingly introduced mandatory minimum
prison terms. 68 The Part III analysis suggests that the legislature oversteps the bounds of good decisionmaking when it does this, by taking for
itself the rulemaking (Decision 2) and the punishment amount and
method decisions (Decisions 5 and 6). These decisions require attention
to the details of the case circumstances and a familiarity with the facts of
the individual offender that the legislature cannot have. While
mandatory minimums might have been justified in a nonguidelines
world-one might argue that they have had a positive, albeit heavyhanded, effect in increasing uniformity-once a guidelines system is put
in place, the need for mandatory minimums disappears. Moreover,
mandatory minimums seriously interfere with the rational proportionality
among offenses that the guidelines seek to introduce. The continuing
use of mandatory minimums means a sentencing system that regularly
fails to distinguish among meaningfully different cases and that systematically does injustice.
B. FederalSentencing Reform Act of 1984
The federal Sentencing Reform Act of 198469 accomplished, among
other things, the creation of the United States Sentencing Commission.
The authority and responsibilities the Act gave to the Commission are in
large part consistent with what Part III finds suitable. The legislature itself set the balance between uniformity and flexibility (by allowing a sentencing range for each guideline category but requiring that the top of
the range not exceed the bottom by more than 25% or six months,
whichever is greater), 7° but generally left it to the Commission to articulate the specific rules by which the purposes of sentencing were to be
achieved.7' The Act also required the Commission to report its work for
review by the legislature, 7 2 again, consistent with Part III conclusions.
68. See Schulhofer, supra note 2, at 201 (discussing increase in federal mandatory
minimum sentence provisions since mid-1980s).
69. See Pub. L. No. 98-473, 98 Stat. 1987 (codified as amended in scattered sections of
18 U.S.C. & 28 U.S.C. (2000)).
70. 28 U.S.C. § 994(b). Of course, this aspect of the Sentencing Reform Act has been
the subject of much controversy. Many judges, certainly, would have struck the balance
between uniformity and flexibility differently, preferring to allow more flexibility. This is
the result that the Supreme Court has now directed in United States v. Booker, 125 S.Ct.
738, 764 (2005) (Breyer, J., opinion of the Court) (invalidating statutory provision that
made the federal guidelines mandatory). Nothing in the analysis in Part III mandates the
balance struck by the Sentencing Reform Act; it speaks only to the distinct issue of who
should decide.
71. See 28 U.S.C. § 995(a)(1) (authorizing the Commission to "establish general
policies and promulgate such rules and regulations for the Commission as are necessary to
carry out the purposes of this chapter").
72. See id. § 9 9 4 (p) (requiring the Commission to submit to Congress amendments
and modifications to amendments of the guidelines and providing that "[s]uch an
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1153
On the other hand, the Act did delegate much of the policymaking
(Decision 1) to the Commission, in terms not entirely consistent with Part
III's suggestions. The Act does not set the relative priority among conflicting sentencing goals, apparently believing that broad policymaking
would benefit from expertise and continuing study that the legislature
itself could not as effectively provide. 73 Instead, it requires the appointment of Commissioners with expertise7 4 and directs the Commission to
take account of existing research, to conduct its own research, and to
support that of others. 75 While this does not match the allocation of authority recommended in Part III, one can perhaps understand the legislative reasoning: The Commissioners might be better positioned to make
decisions about how to choose among sentencing purposes when alternative purposes conflict.
Unfortunately, this aspect of the legislative delegation turned out
badly, and Congress failed to exercise the veto power it held when the
guidelines were presented to it for promulgation. The Commission in
fact failed to do what the legislature directed in its policymaking delegation. It failed to determine the purposes, or their interrelation with one
another, that would guide the formulation of guidelines. Instead, it
opted for what it saw as the politically expedient approach of simply fol-
amendment or modification shall be accompanied by a statement of the reasons
therefor").
73. 18 U.S.C. § 3553(a) (2) sets forth broad purposes of sentencing:
(A) to reflect the seriousness of the offense, to promote respect for the law, and
to provide just punishment for the offense;
(B) to afford adequate deterrence to criminal conduct;
(C) to protect the public from further crimes of the defendant; and
(D) to provide the defendant with needed educational or vocational training,
medical care, or other correctional treatment in the most effective manner.
74. See 28 U.S.C. § 991 (a) (describing appointment procedures); id. § 991 (b) (1) (C)
(describing one purpose of the Commission as to establish sentencing policies and
practices that "reflect, to the extent practicable, advancement in knowledge of human
behavior as it relates to the criminal justice process").
75. Id. § 994(d)-(f) (requiring that the Commission determine relevance of certain
factors to sentencing and include those factors in guidelines and policy statements); id.
§ 994(m) (instructing the Commission to determine average lengths of sentences imposed
and served for categories of cases); id. § 994(o) (providing that the Commission shall
review its work in consideration of comments and data it receives and that various agencies
shall provide the Commission with assessments of its work); id. § 994(w) (requiring
sentencing courts to provide written reports of sentences and relevant information to the
Commission, which shall make these records, analyses of the records, and data files based
on the records available to Congress and the Attorney General); id. § 995(a)(12)
(authorizing the Commission to establish a "research and development program" to serve
as a "clearinghouse and information center" for information on federal sentencing
practices); id. § 995(a) (13)-(16) (authorizing the Commission to systematically collect and
disseminate research and information on sentencing process, sentences imposed,
relationship of imposed sentences to sentencing purposes, and effectiveness of imposed
sentences).
1154
COLUMBIA LAW REVIEW
[Vol. 105:1124
lowing mathematical averaging of past sentencing practices. 76 This approach was thought attractive because it would allow the Commission to
counter any criticism by claiming that its guidelines did little more than
continue existing sentencing practice. The Commission failed to rely
upon, or even to inform itself of, existing research. 77 The ultimate effect
was a set of guidelines with no principled basis. Thus, no explanation
could be given for any sentence other than to say that the sentence was
consistent with the mathematical average of previous sentences. It is not
hard to imagine that for the sentencingjudge who must face the offender
being sentenced, this is not an explanation that carries much credibility
or majesty.
C. The Feeney Amendment to the Federal PROTECT Act: As Proposed, as
Enacted
The original proposal of the federal PROTECT Act would have required the United States Sentencing Commission to forbid judicial downward departures from the guideline sentence unless expressly provided
and, further, would have placed a moratorium on promulgating new
grounds for downward departures. 78 In other words, the legislature
sought to take back for itself much of its delegation of rulemaking authority (Decision 2). The motivation for the reform was a purported concern for widespread judicial subversion of the guidelines through downward departures, 79 but the claim was simply not supported by the
evidence. 80 The unfortunate truth may have been an instance of the
kind of political law-and-order posturing that the Sentencing Reform Act
of 1984 had so selflessly tried to put beyond easy reach of politicians.
76. Dissenting View of Commissioner Paul H. Robinson to the Promulgation of
Sentencing Guidelines by the United States Sentencing Commission, 52 Fed. Reg. 18,121,
18,122 (1987).
77. Id. at 18,121-22.
78. Vinegrad, supra note 5, at 311, 314-16.
79. See, e.g., id. at 316-19 (discussing proponents' reliance on the unsupported
notion that downward departures had become routine and "[c]ongressional disdain for
the exercise of sentencing discretion by federal district judges").
80. In fiscal year 2002, 65% of cases resulted in sentences within the guideline range,
and approximately 34% of cases involved downward departures. However, slightly over
half of the downward departure cases were substantial assistance departures (requested by
prosecutors), leaving only 16.8% of cases as other downward departures. U.S. Sentencing
Commission, 2002 Sourcebook of Federal Sentencing Statistics 53 tbl.26 (2003). Only
21.3% of these non-substantial assistance downward departure cases, or approximately 4%
of all cases, were for "general mitigating circumstances," id. at 52 tbl.25, the type of cases
that supporters of the Feeney Amendment point to as an example of abuse. In all other
downward departure cases, a specific reason such as deportation or age was given. Id.
Downward departures pursuant to a plea agreement (18.7% of non-substantial assistance
downward departure cases) were almost as frequent as downward departures for general
mitigating circumstances at 21.3%. Id. In fiscal year 2002, the government only appealed
183 issues in 110 cases. Id. at 109 tbl.58. Only twenty-six of those issues, or 14.2%, involved
U.S. Sentencing Guidelines Manual § 5K2.0 departures, amounting to only 0.1% of the
over 20,000 departure cases. Id.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1155
Thus, the sad episode may illustrate the wisdom of the Sentencing Reform Act of 1984 in its effort to distance the criminal punishment process
from the rough-and-tumble dynamics of politics we discussed in Part 111.81
After objections from many quarters, the Feeney Amendment was
scaled back, but only to reduce the breadth of the Amendment's effects,
not its objectionable nature. The prohibitions of downward departures
and of the introduction of new grounds for downward departures were
limited to a list of child abduction, pornography, and sexual assault
82
offenses.
Perhaps of greatest concern in the Feeney Amendment is that its
sponsors fundamentally misperceive the nature of downward departures
and by their Amendment have helped spread that misperception. The
sponsors apparently perceive any downward departure as suspect, a
clearly bad thing, something to be "substantially reduced," 83 and a corruption of the process. The Amendment's supporters criticize downward
departures by complaining that laws on criminal sentencing should be
"faithfully, fairly, and consistently enforced,"8 4 as if a downward departure was a breach of that faith. This view of downward departures is particularly peculiar because, in their current form, the guidelines are constructed to actually expressly authorize judges to downwardly depart in
certain circumstances, so judges are not in reality departing from the
85
guidelines at all when they follow one of these Commission invitations.
The truth is that downward departures can be a sign of health in a
guideline system and their absence a sign of dysfunction. No set of guidelines can anticipate and articulate rules for every conceivable variation in
human conduct, situation, and personality; we would not want a guidelines system to try to do so because there is an infinite variety of possibilities. The Sentencing Reform Act of 1984 made clear that the goal was to
strike a proper balance between the goal of uniformity and the goal of
individualizing justice. It would be anathema to the Act to ignore relevant factors simply because they could not be articulated beforehand. An
absence of downward departures would clearly signal that criminal sen81. See supra text accompanying notes 22-23.
82. PROTECT Act of 2003, Pub. L. No. 108-21, 117 Stat. 650 (codified as amended in
scattered statutes of 18 U.S.C. and 28 U.S.C. (2000)).
83. Id. § 401(m).
84. Memorandum from Attorney General John Ashcroft, to All Federal Prosecutors,
Department Policy Concerning Charging Criminal Offenses, Disposition of Charges, and
Sentencing 1 (Sept. 22, 2003), available at http://www.nacdl.org/public.nsf/2cdd02b415
ea3a64852566d6000daa79/departures/$FILE/AshcroftCharging%2OMemo.pdf
(on file
with the Columbia Law Review) (directing federal prosecutors to acquiesce to downward
departures only in limited circumstances because the PROTECT Act "emphatically
reaffirms Congress' intention that the Sentencing Reform Act and the Sentencing
Guidelines be faithfully and consistently enforced").
85. U.S. Sentencing Guidelines Manual § 5K2.0 (2003) (describing circumstances in
which "[t]he sentencing court may depart from the applicable guideline range").
1156
COLUMBIA LAW REVIEW
[Vol. 105:1124
tencing had gone awry, and that justice was being sacrificed to
uniformity.
To tie this to the analysis in Part III, we noted in relation to Decision
5, determining punishment amount, that a sentencing commission would
probably vary the level of control overjudicial decisionmaking by making
guidelines more or less binding as the situation requires. The Commission's formal recognition of certain grounds for downward departure and
the guidelines' recognition that downward departure is also authorized in
other situations "not adequately taken into consideration by the Sentencing Commission in formulating the guidelines"8 6 perfectly illustrate this
dynamic in action. As Part III suggests, such a de facto delegation of
power to judges is necessary for an effective sentencing system; it is not a
breach of faith with criminal sentencing laws. Further, in relation to Decision 2, articulating sentencing rules, the Commission ought to establish
a feedback loop with the frontline users, the judges, to help refine and
adjust the guidelines to improve their formulation and application. The
downward (and upward) departure authority of judges is an important
part of that feedback loop. If a clear pattern develops in the use of downward departures, the Commission is alerted to consider whether some
refinement is needed. To shut out this feedback loop is to invite ignorance and stagnation in guideline development.
Congress showed great self-knowledge and self-restraint in the Sentencing Reform Act of 1984, but seems to have lost its way a bit when it
tried to undo itself in the PROTECT Act.
D. Blakely v. Washington
In Blakely v. Washington, the defendant's sentence for second degree
kidnapping with a firearm was set above the "standard range" of fortynine to fifty-three months of the state's sentencing guidelines, to a sentence of ninety months, based upon the judge's finding after a post-trial
hearing that the offense was committed with "deliberate cruelty,"8 7 a statutorily enumerated ground for upward departure from the guideline
standard range in domestic violence cases.8 8 The Court followed its reasoning in Apprendi v. New Jersey89 and held that "every defendant has the
right to insist that the prosecutor prove to ajury all facts legally essential
to the punishment,"90 including, in this instance, the finding of "deliberate cruelty" that brought the defendant an additional thirty-seven months
of prison.
Blakely appears to constitutionalize a rule that both factfinding and
judgment-making (Decisions 3 and 4) must be left to the jury. While
86.
87.
88.
89.
90.
Id. § 5K2.0 (a)(1)(B).
Blakely v. Washington, 124 S. Ct. 2531, 2537 (2004).
Id. at 2535.
530 U.S. 466, 483-84 (2000).
Blakely, 124 S. Ct. at 2543.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1157
there may be some question about whether such a rule merits constitutional status, as a matter of policy it reflects a view consistent with the
optimal decisionmaking analysis offered in Part III. Indeed, the "deliberate cruelty" determination in Blakely is a textbook example of Decision 4
judgment-making (as compared to Decision 3 factfinding), for which the
advantage ofjury decisionmaking is at its clearest advantage over that of a
judge, for the decision calls for an exercise of normative judgment rather
than simply a reconstruction of past events.
One can imagine difficulties in the application of the Blakely rule:
Which facts relevant to sentencing must be determined by a jury? Must
the jury determine such things as an offender's amenability to treatment,
or his likely future dangerousness? The amount-method distinction reflected in the difference between Decision 5 and Decision 6 may help
resolve this difficulty. Setting the amount of punishment, Decision 5, is
the decision for which we have the greatest fear of government abuse and
the greatest need to control governmental intrusion in our lives. Thus, it
seems entirely appropriate that the jury should be responsible for the
factfinding and judgment-making decisions (Decisions 3 and 4) that underlie the determination of punishment amount (Decision 5). The determinations that underlie the optimum-method-of-punishment decision,
Decision 6, may be similar in some respects to those underlying Decision
5, but here, we do not have the same fundamental interests at stake. If
the method decision is limited in its overall punitive bite by the previously
made amount decision, our concerns for undue governmental intrusion
and abuse have already been addressed. Thus, reliance upon a judge's,
rather than ajury's, factfinding andjudgment-making seems unobjectionable and indeed has the advantages described in Part III.F above.
The failure to see the decisionmaking advantages of ajury over those
of a judge in sentencing matters has led many people to view the Court's
Blakely decision as a thinly veiled judicial attack on the sentencing guidelines generally, and the federal guidelines in particular, demonstrating
the hostility that has existed ever since judges had their broad discretion
reduced. 9' But the decision does have a rational basis in the optimal
91. See, e.g., United States v. O'Meara, 895 F.2d 1216, 1221-23 (8th Cir. 1990)
(Bright, J., dissenting in part) ("This case opens the window on the sometimes bizarre and
topsy-turvy world of sentencing under the Guidelines. ... I cannot help but feel that we
have lost something in this substitution of technical proficiency for the thoughtful exercise
of discretion by the federal judiciary."); Frank S. Gilbert, A Probation Officer's Perception
of the Allocation of Discretion, 4 Fed. Sent'g Rep. 109, 109 (1991) ("A well-known and
highly respected jurist recently observed that his sentencing role has been relegated to that
of a 'notary public.' ");Jack B. Weinstein, A Trial Judge's Second Impression of the Federal
Sentencing Guidelines, 66 S. Cal. L. Rev. 357, 363 (1992) ("[T]bere is considerable
justification for believing that guideline sentencing will continue to undermine federal
criminal adjudication in more subtle and insidious ways."); John S. Martin, Jr., Editorial,
Let Judges Do Their Jobs, N.Y. Times, June 24, 2003, at A31 (citing guidelines as a reason
for his retirement from federal judiciary and stating, "I no longer want to be part of our
unjust criminal justice system").
1158
COLUMBIA LAW REVIEW
[Vol. 105:1124
decisionmaker analysis, and there may be no other hidden agenda. The
Court could cite the social science studies noted in Parts III.C & D in
support of its decision.
Of course, the Court itself has raised some suspicion that the basis
for Blakely is more logical than political. After finding in Booker9 2 that the
federal sentencing guidelines violated the constitutional rule announced
in Blakely, the Court did not construct a remedy that advanced jury
factfinding, as it could have done by requiring a bifurcated procedure
with a post-conviction jury factfinding hearing for sentencing. 93 (Of
course such sentencing hearings would be rarer than trials.) Instead, the
Court abolished the need for any jury basis for sentencing issues! 9 4 The
Court converted its announced constitutional right to jury factfinding
and jury judgment-making (consistent with cells D3 & D4) into judicial
authority to engage in policymaking, rule articulation, factfinding, and
judgment-making-decisions for which judges are not the most effective
decisionmaker. (See cells El, E2, E3 & E4.) The supposed constitutional
need forjury decisionmaking for sentencing has been remedied by giving
judges greater decisionmaking authority. It is hard not to see this contorted remedy as undermining the legitimacy of the Court's justification
for creating the constitutional rule.
E. Jury Sentencing
In six states,juries select an appropriate sentence from within a statutory range of punishment.9 5 There is renewed interest in promoting general jury sentencing more widely.9 6 Juries may be good-perhaps better
than judges-at making reliable factual determinations and judgments
that accurately capture shared community intuitions ofjustice (Decisions
3 and 4), but this does not justify giving juries general sentencing authority (thereby including Decisions 1, 2, 5, and 6) because the exercise of
such sentencing authority requires different attributes that juries do not
92. United States v. Booker, 125 S. Ct. 738 (2005).
93. Id. at 759-64 (Breyer, J., opinion of the Court).
94. Id. at 764 (Breyer, J., opinion of the Court).
95. See King & Noble, supra note 9, at 886-87.
96. See, e.g., Hoffman, supra note 8, at 956 (suggesting that "jurors are better than
judges at imposing appropriate criminal punishment"); Iontcheva, supra note 14, at 314
(urging legislatures to reintroduce jury sentencing); King & Noble, supra note 9, at 886-87
(discussing recent "renewed interest in the appropriate role for juries in sentencing").
The statutory ranges from which juries choose typically are quite broad. Id. at 892. In
Kentucky, juries involved in felony sentencing pick from the statutory ranges of 20 to 50
years or life imprisonment for a Class A felony, 10 to 20 years for a Class B felony, 5 to 10
years for a Class C felony, and 1 to 5 years for a Class D felony. Ky. Rev. Stat. Ann.
§ 532.060(2) (Michie 1999). In Arkansas and Virginia, juries select from similarly wide
ranges. See Ark. Code Ann. § 5-4-401 (a) (Michie 1997); Va. Code Ann. § 18.2-10 (Michie
2004). Even though Virginia has voluntary judicial sentencing guidelines, the narrower
guideline ranges are not revealed to the jury. King & Noble, supra note 9, at 893.
Arkansas also has sentencing guidelines, but again, only forjudges. Id. Kentucky does not
have sentencing guidelines. Id.
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1159
possess. As Part III suggests, juries are ill-equipped to make policy decisions regarding the proper purposes of punishment, as called for in Decision 1, or to determine how those purposes are best operationalized, as
called for in Decision 2 and applied in Decision 5. Nor are juries aware
of the strengths, weaknesses, applicability, or availability of the various
sentencing alternatives, as Decision 6 requires. Beyond these informational, experiential, and educational lapses, jury sentencing (like broadly
discretionary judicial sentencing) would produce serious disparities in
the treatment of meaningfully similar cases. Without articulated rules,
different juries may come to different results in sentencing similar
offenders.
F. State Sentencing Guidelines Systems
Considerable diversity exists among state guidelines systems, but one
can nonetheless draw some general conclusions about how they, as a
group, are different from the federal guidelines, and some of these differences have implications for determining who should make what sentenc97
ing decisions.
For example, in most state systems judges retain greater discretion
than is allowed federal judges, and in some state systems judges are allowed much more discretion. 98 Many writers aggressively support less
binding guidelines and greater judicial discretion, 99 but Part III suggests
that the issue is somewhat more complex than such a unidimensional
view would permit. Whether discretion is a good or bad thing depends
upon the particular nature of the matter on which the discretion is given.
Discretion to determine the sentencing purpose to be followed in a particular case (Decision 1) is seriously problematic for all the reasons dis97. Many of the differences between the federal guidelines and state guidelines do
not relate to who makes the sentencing decisions, including, for example, the extent to
which the guidelines are constrained by the prison space currently available, the particular
sentencing purpose the guidelines adopt, and the variety of offender characteristics of
which the guidelines take account. See Richard S. Frase, State Sentencing Guidelines: Still
Going Strong, 78 Judicature 173, 174-77 (1995) [hereinafter Frase, State Sentencing
Guidelines] (contrasting goals and priorities of state guidelines with federal approach).
For a general review of state sentencing guidelines, see Richard S. Frase, Is Guided
Discretion Sufficient? Overview of State Sentencing Guidelines, 44 St. Louis U. L.J. 425,
426 (2000) (examining the "evolution, successes, and failures" of state sentencing
guidelines); Michael Tonry, Sentencing Commissions and Their Guidelines, 17 Crime &
Just. 137, 138-42 (1993) (discussing the survival of state sentencing commission systems);
Tonry, Judge Frankel, supra note 12, at 714 (canvassing state sentencing commissions that
"have successfully established and implemented responsible sentencing policies").
98. See, e.g., Frase, State Sentencing Guidelines, supra note 97, at 175. Obviously, any
kind of sentencing system must retain some discretion for sentencing judges. The Part III
analysis of Decisions 5 and 6 acknowledges this fact.
99. See, e.g., id. (discussing judicial discretion of state guidelines and criticizing
federal guidelines as "uniquely and unnecessarily rigid"); Tonry, Judge Frankel, supra note
12, at 716, 721-22 (criticizing lack of judicial discretion in federal guidelines).
1160
COLUMBIA LAW REVIEW
[Vol. 105:1124
cussed above. 10 0 It has judges making policy decisions that a commission
could more expertly make and creates disparity among judges as differentjudges follow different purposes. Discretion to operationalize a guiding purpose into a sentencing rule (Decision 2) suffers the same objections. Discretion to determine the facts and to make the judgments upon
which sentencing is to be based (Decisions 3 and 4) may be somewhat less
objectionable, but as Part III suggests (and Blakely v. Washington may require), such discretion might be better exercised byjurors. On the other
hand, it may be very true to say that judges ought to have much discretion
to determine the method of punishment (Decision 6), and perhaps even
in determining the exact amount of punishment through application of
the guidelines (Decision 5), because the guidelines must leave some
room for discretion as not every relevant factor can be anticipated and
articulated. 101
This last point highlights another common difference between state
and the federal guidelines: The state systems are more likely to support
greater use of "intermediate sanctions," meaning noncustodial sanctions. 10 2 Part III's analysis very much supports the state practice in this
regard. As long as the amount of punishment is determined under articulated rules (Decision 2) as applied to the facts of the case (Decisions 3
and 4),judges ought to have wide discretion to determine the method of
punishment (Decision 6). Such discretion does not risk unjustified disparity and is well informed by the special circumstances of each case and
the available punishment options, which only the sentencing judge with
her local knowledge and expertise is likely to fully understand.
One final example of state-federal difference concerns the makeup
of the sentencing commission. Many state commissions have broader
representation than is called for in the federal commission. 10 3 While the
federal commission typically has sentencing judges and experts of one
kind or another (criminal lawyers, professors, or corrections officials are
typical), some state systems have and others expressly require "public
members" and/or legislators, who may have no expertise regarding criminal justice or punishment issues.
100. See supra Part lI.A.
101. See supra Part II.F.
102. See, e.g., Frase, State Sentencing Guidelines, supra note 97, at 174 (noting that
"guidelines states have increasingly encouraged judges to employ 'intermediate sanctions,'
or non-custodial penalties more intensive than simple probation" while "the federal
guidelines have done very little to promote" use of such sanctions).
103. Compare Minn. Stat. Ann. § 244.09 (West 2003) (directing that sentencing
guidelines commission shall include, among other people, judges, attorneys, and three
public members), and Wash. Rev. Code Ann. § 9.94A.860 (West 2003) (directing that
sentencing guidelines commission shall include, among other people, a law enforcement
officer, elected officials, and four members of the public, one of whom is a crime victim or
"victim's advocate"), with 28 U.S.C. § 991 (a) (2000) (giving no more direction on the
composition of the Commission than limiting the number of federal judges and the
number of members of the same political party).
2005]
MATCHING DECISIONMAKER TO DECISION NATURE
1161
If one has a monolithic view of the sentencing decision, this greater
representativeness may seem appealing. The use of "public members"
has the feel of the legitimacy that juries bring; the inclusion of legislative
members seems to give voice to the important policymaking role we give
to legislatures. But once one sees, as Part I suggests, that the sentencing
decision is not a monolith but rather a series of decisions, each of which
is best allocated to a different decisionmaker, then the broad representation of some state systems seems a bit misguided. Rather than having a
legislator on the commission, it is better to have the legislature take full
control over those policymaking decisions that benefit from the special
representative nature of that body, as the Decision 1 discussion suggests.
The involvement of members of the public ought to be directed to those
aspects of the decisionmaking that benefit from those special characteristics in the context of jury factfinding and judgment-making (Decisions 3
and 4). Given the nature of the decisions that the commission must
make-rulemaking, and some delegated policymaking (Decision 2 and
part of Decision 1)-the commission ought to have people who have the
training and expertise that those decisions require.
CONCLUSION
A failure to see that the sentencing decision has several discrete parts
requiring different decisional qualities often leads to reforms that may
make sense as to one decision but not others. Juries may indeed be superior factfinders and judgment-makers-support for the Blakely rule-yet
would perform poorly in taking on broader sentencing duties, as jury sentencing systems commonly have them do. Legislatures may indeed be
preferable for setting broad policy directions, but they are poor in making most other sentencing decisions-hence the dysfunction of
mandatory minimum sentences. Judges may be the optimal choice to decide the method of punishment in the individual case-prison, fine, drug
treatment, house arrest, and so forth-but it does not follow that judges
are the best decisionmaker to decide the guiding purposes that should
drive the punishment system, as the proponents of broad judicial sentencing would have. The optimal sentencing decisionmaking is one that
matches each aspect of that process with the decisionmaker who is best
suited to the task.